New insights into PTEN

Phosphatase and tensin homolog (PTEN) was identified in 1997 as the relevant gene in a region of chromosome 10 that is often lost in late-stage human tumors, especially those of the brain, prostate and endometrium (Li et al., 1997; Steck et al., 1997). Soon after, it was discovered that PTEN demonstrates phosphatase activity against the phospholipid products of PI3-kinase activity, namely phosphatidylinositol (3,4,5)-trisphosphate [PtdIns(3,4,5)P₃], as well as other 3′ phosphorylated phosphoinositides, albeit to a lesser extent (Maehama et al., 2001). PTEN can also dephosphorylate protein substrates such as FAK (PTK2) (Tamura et al., 1999), as well as itself (Raftopoulou et al., 2004). The relative contributions of the lipid phosphatase activity and protein phosphatase activity, as well as additional phosphatase-independent functions, to the roles of PTEN in development and normal physiology, as well as in pathological states such as diabetes and cancer, are only just beginning to be unraveled. This Commentary focuses on recent insights into PTEN, in particular those that shed light on the regulation of its levels and activity, its role in human diseases, and novel downstream functions independent of PI3-kinase.

PTEN was initially assumed to be constitutively expressed, but several transcription factors have now been shown to bind directly to the PTEN promoter and regulate PTEN expression (Fig. 1). Following sequencing of the promoter, Virolle et al. noticed potential binding sites for early growth regulated transcription factor 1 (EGR1) (Virolle et al., 2001) and showed that PTEN is upregulated by EGR1 in response to radiation treatment (Virolle et al., 2001), and by IGF2 – as part of a negative-feedback loop (Moorehead et al., 2003). EGR1 also upregulates PTEN following treatment with the phosphatase inhibitor calyculin A, which might contribute to the apoptotic effects of this agent (Okamura et al., 2005). EGR1 levels strongly correlate with PTEN levels in a cohort of non-small cell lung cancer (NSCLC) tumors, and upregulation of these two proteins is associated with poor prognosis (Ferraro et al., 2005). An association between EGR1 and PTEN levels in thyroid cancers is not as clear (Di Loreto et al., 2005); PTEN is therefore likely to be regulated by other factors as well.

Indeed, Patel et al. have shown that peroxisome proliferator-activated receptor γ (PPARγ) also regulates PTEN transcription (Patel et al., 2001). Activation of PPARγ by its selective ligand, rosiglitazone – an anti-diabetic drug from the thiazolidinedione class – leads to binding of PPARγ to the PTEN promoter at two sites (PPRE1 and PPRE2) and to the consequent upregulation of PTEN in normal and cancer cells in humans. This is accompanied by a decrease both in phosphorylation of PKB (AKT) and in cell proliferation, leading the authors to speculate that PPARγ agonists might be beneficial for the treatment of cancers in which PTEN is still functional. For example, rosiglitazone increases PTEN expression and reduces migration in the hepatocarcinoma cell line BEL-7404 (Zhang, W. et al., 2006). Han and Roman showed growth suppression by rosiglitazone in NSCLC cells (Han and Roman, 2006). This effect has a PPARγ- and PTEN-dependent component but also a PPARγ-independent component via increased phosphorylation of AMP-activated protein kinase, which suppresses growth by inactivating the mammalian target of rapamycin (mTOR, also known as FRAP1).

p53 can also transcriptionally upregulate PTEN by binding to its promoter (Stambolic et al., 2001). A complex interplay exists between p53 and PTEN, which regulate the levels and activities of each other in several ways. For example, PTEN is required for p53-mediated apoptosis in immortalized mouse embryonic fibroblasts (MEFs). Indeed, p53^–/– MEFs display only ∼30% of the PTEN protein levels seen in p53^+/+ cells (Wang et al., 2005).

Another protein that modulates growth factor actions, human sprouty homolog 2 (SPRY2), has recently been shown to mediate its anti-proliferative effect by the regulation of PTEN. Over-expression of SPRY2 upregulates PTEN mRNA and protein levels (but decreases phosphorylation of PTEN at S370 and S385, and its stability), and the ability of SPRY2 to restrain proliferation is abolished in PTEN-negative cells (Edwin et al., 2006).

Resistin, a cytokine involved in inflammation and insulin resistance, increases PTEN expression in human aortic vascular endothelial cells by activating p38 MAPK and activating transcription factor 2 (ATF2), which leads to decreased activation of PKB and of its target, endothelial nitric oxide synthase (eNOS, also known as NOS3). This provides a potential mechanism for resistin-mediated impairment of insulin signaling and of its role in cardiovascular diseases (Shen et al., 2006). Furthermore, phytoestrogens such as genistein (in soy), resveratrol (in red wine) and quercetin (in fruit and vegetables) also lead to increased PTEN expression, with a concomitant decrease in phospho-PKB levels and increase in levels of the CDK inhibitor p27 (Waite et al., 2005). Epidemiological data suggest that phytoestrogen consumption can protect against cancer (Park and Surh, 2004; Verheus et al., 2007). Upregulation of PTEN expression might therefore be one basis for these beneficial effects. Indeed, mammary glands of rats fed with soy protein or a genistein-supplemented diet display higher PTEN levels and a higher apoptotic index (Dave et al., 2005). The same is true for MCF-7 cells cultured in serum from such rats (Dave et al., 2005). Similarly, indole-3-carbinol, a phytochemical derived from cruciferous vegetables such as broccoli, upregulates PTEN expression in the cervical epithelium in a mouse model for cervical cancer (Qi et al., 2005).

Most of the data so far have demonstrated positive regulation of PTEN transcription. Negative regulation of PTEN expression has also been shown, however (Fig. 1). Mitogen-activated protein kinase kinase-4 (MKK4) inhibits PTEN transcription by activating NFκB, which binds to a site in the PTEN promoter (Xia et al., 2007). Transforming growth factor (TGF)β also decreases PTEN transcription in pancreatic cancer cells (Chow et al., 2007) and in mesangial cells (Mahimainathan et al., 2006). Likewise, the proto-oncogenic transcription factor JUN suppresses PTEN expression by binding to a variant AP-1 site in the PTEN promoter (this site is named PF1), and an inverse correlation between JUN and PTEN levels has been found in a panel of human tumor cell lines (Hettinger et al., 2007).

PTEN levels are also modulated by changes in protein stability due to interactions with other proteins. Through its C-terminal PDZ-domain-binding motif, PTEN interacts with various proteins that contain PDZ motifs. Wu et al. showed that PTEN interacts with members of the membrane guanylate-kinase inverted (MAGI) family, such as MAGI2, which results in its stabilization (Wu, X. et al., 2000; Wu, Y. et al., 2000). More recently, the ubiquitously expressed cytoskeletal protein vinculin was shown to modulate this interaction. Vinculin-null cells have undetectable PTEN protein despite normal mRNA levels (Subauste et al., 2005). Subauste and colleagues have demonstrated that vinculin is required to maintain β-catenin–MAGI2 interactions at adherens junctions and thereby limit ubiquitin-mediated degradation of PTEN (Subauste et al., 2005). Recently, Valiente et al. showed that binding of PTEN to the microtubule-associated serine/threonine (MAST) kinases MAST3 and syntrophin-associated serine/threonine kinase (SAST, also known as MAST1) also stabilizes PTEN and facilitates its phosphorylation by these kinases (Valiente et al., 2005). PTEN-p53 interactions, by contrast, promote caspase-mediated degradation of PTEN (Tang and Eng, 2006b).

PTEN activity is heavily regulated by phosphorylation, acetylation, oxidation and control of its localization (Fig. 2). Below, we summarize some of the findings in this field; further details can be found in several excellent reviews published elsewhere (Baker, 2007; Gericke et al., 2006; Gil et al., 2007; Leslie, 2006; Vazquez and Devreotes, 2006).

PTEN is regulated by phosphorylation of a cluster of serine and threonine residues in its C-terminus (Fig. 2A). Phosphorylation of these residues is not strictly required for activity but stabilizes PTEN; alanine substitutions at these phosphorylation sites generally lead to higher catalytic activity but also to rapid degradation of PTEN in cells. Casein kinase 2 (CK2, also known as CSNK2) has previously been implicated in phosphorylation of some of these sites (Torres and Pulido, 2001). More recent evidence suggests that it might also prime some sites for phosphorylation by glycogen synthase kinase 3β (GSK3β). CK2 mainly phosphorylates S370 and S385, whereas GSK3β targets S362 and T366. The involvement of GSK3β could be part of a negative-feedback loop that regulates PTEN and PI3-kinase activity (Al-Khouri et al., 2005).

Neither CK2 nor GSK3β affects phosphorylation at S380 (Al-Khouri et al., 2005), but glioma tumor suppressor candidate region 2 (GLTSCR2, also known as PICT-1) was recently shown to interact with PTEN, promote its phosphorylation at S380 and upregulate its levels (Okahara et al., 2006; Yim et al., 2007).

Odriozola et al. have studied activity and localization of several C-terminal mutants of PTEN (Odriozola et al., 2007). They propose that the phosphorylated C-terminal tail (residues 380-385) interacts with the C2 and phosphatase domains in PTEN, serving as a pseudosubstrate and therefore causing auto-inhibition. According to this model, dephosphorylation of S385 followed by dephosphorylation of other residues in its vicinity leads to a more open conformation and unmasks the catalytic pocket, which leads to increased membrane affinity and increased PTEN activity (Fig. 2B). Additionally, there is evidence that RhoA-associated kinase (ROCK) phosphorylates S229, T232, T319 and T321 in the C2 domain to activate PTEN and target it to the membrane in chemoattractant-stimulated leukocytes (Li et al., 2005) (Fig. 2A). Similarly, the findings of Papakonstanti et al. place RhoA and ROCK upstream of PTEN (Papakonstanti et al., 2007). The mechanistic consequences underlying ROCK-mediated phosphorylation and activation of PTEN are not understood. Unexpectedly, the PI3-kinase catalytic subunit p110δ was found to inactivate PTEN through a pathway involving RhoA and ROCK, and increased tyrosine phosphorylation of PTEN, although the mechanism by which this occurs remains to be determined (Papakonstanti et al., 2007).

Acetylation is another mechanism that appears to regulate PTEN activity (Okumura et al., 2006) (Fig. 2A). The histone acetyltransferase p300/CBP-associated factor (PCAF) interacts with PTEN and acetylates lysines 125 and 128 in response to growth factors. These residues are located within the catalytic cleft of PTEN and are essential for PtdIns(3,4,5)P₃ specificity; consequently, PCAF functions as a negative regulator of PTEN.

PTEN activity can also be downregulated by reactive oxygen species (ROS), which oxidize the catalytic site cysteine residue, C124, to form an intramolecular disulfide bond with C71. Initially thought to be merely by-products of metabolism, ROS are now known to play a role in cell signaling. Using two neuroblastoma cell lines, one that can produce ROS and one that cannot, Seo et al. demonstrated that generation of ROS and subsequent inactivation of PTEN is required for insulin-mediated activation of PKB (Seo et al., 2005) (Fig. 2A).

Localization of PTEN plays an important role in the regulation of its activity (Fig. 2B, C), as mentioned above. The best described function of PTEN is the dephosphorylation of PtdIns(3,4,5)P₃ at the plasma membrane. To access its target, PTEN needs to be recruited to the membrane. The C2 domain (Das et al., 2003), the phosphatase domain (Das et al., 2003) and the PtdIns(4,5)P₂-binding domain (PBD) (Downes et al., 2007; Vazquez and Devreotes, 2006) of PTEN are crucial for its binding to lipids. Additionally, PTEN is recruited to the membrane by interaction with proteins through its PDZ-binding motif (see above). Among these are MAGI1b, MAGI2, MAGI3, MAST3 and SAST. PTEN also interacts with NHERF1 (EBP50) and NHERF2 (Na⁺/H⁺ exchanger regulatory factor) adaptor proteins through its PDZ-domain-binding motif, and PTEN and NHERF proteins are found in a ternary complex with the platelet-derived growth factor receptor (PDGFR). Activation of PI3-kinase after PDGFR stimulation is prolonged in NHERF1^–/– MEFs and when NHERF2 is knocked down by RNAi. This indicates that NHERF proteins normally recruit PTEN to PDGFR to restrict the effects of PI3-kinase activation (Takahashi et al., 2006).

Exciting studies using single-molecule total internal reflection fluorescence (TIRF) microscopy in living cells have revealed that PTEN binds to the membrane for several milliseconds, which is sufficient to degrade several PtdIns(3,4,5)P₃ molecules. These also confirm that phosphorylation of the C-terminal tail constrains PTEN conformation, which limits its association with the membrane (Vazquez et al., 2006), supporting the earlier models (e.g. Li et al., 2005; Odriozola et al., 2007).

The existence of nuclear PtdIns(3,4,5)P₃ and PtdIns(4,5)P₂ point to a role for PTEN in the nucleus (Caramelli et al., 1996; Mazzotti et al., 1995), but this function is less clear than its well-established role at the plasma membrane. Whereas some researchers have detected PTEN exclusively in the cytoplasm (Gu et al., 1998), others have found it in both the cytoplasm and nucleus, and loss of nuclear PTEN has been associated with neoplasias and tumor formation (Gimm et al., 2000; Perren et al., 1999), indicating a tumor suppressor function of nuclear PTEN.

Employing different-sized GFP-PTEN fusion proteins, Liu et al. used fluorescence recovery after photobleaching (FRAP) to conclude that PTEN can passively diffuse through nuclear pores (Liu et al., 2005). By contrast, Chung et al. have described bipartite nuclear localization sequences in PTEN that are required for major vault protein (MVP)-mediated nuclear import (Chung et al., 2005). They have also detected different effects of cytoplasmic versus nuclear PTEN: cytoplasmic PTEN decreases phospho-PKB levels, upregulates p27^kip1 and is required for apoptosis, whereas nuclear PTEN downregulates cyclin D1 and phospho-MAPK and is crucial for cell cycle arrest (Chung and Eng, 2005; Chung et al., 2006). Gil et al. have also provided evidence for active transport of PTEN into the nucleus (Gil et al., 2006). Using a series of PTEN mutants, they suggest that multiple nuclear exclusion motifs (in the phosphatase domain, C2-domain and C-terminus) and NLS domains (in the N-terminus) control PTEN localization in a RAN-dependent manner.

The discovery of a Cowden-syndrome-associated lysine mutation in PTEN, K289E, led to further insight into the control of PTEN localization. This mutant retains catalytic activity but fails to accumulate in the nucleus owing to an import defect (Trotman et al., 2007). K289 and other lysines residing in PTEN are monoubiquitylated by the E3 ligase neural precursor cell expressed, developmentally downregulated 4-1 (NEDD4-1) in the cytoplasm, which permits nuclear import. Cytoplasmic monoubiquitylated PTEN can be ubiquitylated further by NEDD4-1 and degraded by the proteasome, can shuttle into the nucleus and back, or can be deubiquitylated in the nucleus, remain nuclear and thus be protected from cytoplasmic degradation (Wang, X. et al., 2007). This connection is illustrated by analyses of PTEN half-life: wild-type PTEN, both cytoplasmic and nuclear, has a t_1/2 of ∼7.5 hours; forced cytoplasmic localization leads to a decrease in stability to t_1/2=4.5 hours; and forced nuclear accumulation increases this to 15 hours (Trotman et al., 2007).

Therefore, the traditional viewpoint that PTEN exerts all of its effects via dephosphorylation of PtdIns(3,4,5)P₃ at the plasma membrane needs to be re-evaluated in light of the overwhelming evidence for the existence and importance of nuclear PTEN. The exact targets and pathways that are affected by nuclear PTEN remain an exciting opportunity for future research.

Many effects of PTEN are due to its ability to dephosphorylate PtdIns(3,4,5)P₃ and thereby antagonize PI3-kinase (Chow and Baker, 2006; Sansal and Sellers, 2004). However, recent evidence indicates that it has effects beyond its function as a phospholipid phosphatase, both in normal and in pathological signaling. For example, an important function of PTEN is likely to be its ability to restrain cell migration. This was shown to be independent of the lipid phosphatase activity of PTEN but dependent on its protein phosphatase activity. Surprisingly, the protein phosphatase activity is required only intramolecularly, and expression of the isolated C2 domain of PTEN, or a catalytically inactive T383A mutant, both inhibit cell migration as effectively as wild-type PTEN (Raftopoulou et al., 2004).

Another clue to the phosphatase-independent functions of PTEN came from studies of p53 regulation. In cells lacking PTEN, p53 levels are significantly reduced owing to decreased stability. Expression of wild-type or phosphatase-dead forms of PTEN increases p53 stability in an MDM2-independent manner (Freeman et al., 2003). This is because the C2 domain of PTEN binds to the C-terminus of p53, which increases its DNA binding and transcriptional activity (Freeman et al., 2003). A similar conclusion was reached by Li et al., who showed that PTEN promotes p53 acetylation through direct association with the histone acetyl transferase p300/CBP in response to DNA damage (Li et al., 2006). Li et al. further demonstrated that PTEN induces p53-acetylation-dependent tetramerization of p53, which promotes interaction with PTEN in a phosphatase-independent manner. This could account for the G1 arrest induced by nuclear-targeted PTEN but not G1 arrest induced by cytoplasmic-targeted PTEN. Liu et al. have already noted the PI3-kinase- and AKT-independent effects of nuclear-targeted PTEN, although they found PTEN phosphatase activity to be required for G1 arrest and for inhibition of soft agar growth (Liu et al., 2005).

Although mutation of one PTEN allele and loss of the second represents the most common mechanism underlying PTEN loss-of-function in human tumors, additional mechanisms have recently been discovered. Methylation of the PTEN promoter is a frequent occurrence in certain types of cancer, such as thyroid cancer (Alvarez-Nunez et al., 2006), melanoma (Mirmohammadsadegh et al., 2006), lung cancer (Marsit et al., 2005) and low grade/secondary glioblastoma mulitforme (GBM) tumors (Wiencke et al., 2007). There does not seem to be a straightforward correlation between PTEN promoter methylation and loss of PTEN expression in all cases analyzed, possibly owing to the small percentage of cells displaying PTEN methylation. Because PTEN methylation correlates with AKT phosphorylation in glial tumors, the most straightforward explanation is that the former causes the latter. However, recent results have shown that AKT activity can influence methylation patterns (Cha et al., 2005), which suggests potential alternative mechanisms.

Additional mechanisms underlying PTEN alterations have also just been uncovered, including promoter mutations in GBM (Tunca et al., 2007), translocations in thyroid cancer (Puxeddu et al., 2005) and Cowden syndrome (CS)-like disease (Yue et al., 2005), and splicing mutations in Bannayan-Riley-Ruvalcaba syndromes (BRRS) (Suphapeetiporn et al., 2006), CS (Agrawal et al., 2005) and sporadic breast cancers (Agrawal and Eng, 2006). Furthermore, a micro-RNA that could target PTEN (miR-21) has been shown to be highly expressed in human cholangiocarcinoma cell lines (Meng et al., 2006) and hepatocellular carcinoma (Meng et al., 2007). Although modulation of miR-21 levels was shown to affect PTEN expression, note that miRs have been suggested to have many potential targets (Bentwich, 2005).

These recent data highlight the variety of genetic and epigenetic alterations that might underlie loss of PTEN in tumorigenesis and could be useful for identifying patients who might be candidates for PI3-kinase-targeted therapies. Interestingly, a germline polymorphism in the 5′ untranslated region (UTR) of the PTEN promoter was associated with patients with type 2 diabetes in a small study of a Japanese cohort (Ishihara et al., 2003). This polymorphism might increase PTEN expression and reduce PI3-kinase signaling in response to insulin (Ishihara et al., 2003), and illustrates the importance of PTEN activity in diseases other than cancer.

Mice systemically lacking both copies of PTEN die early in embryogenesis. Heterozygous mice and chimeric mice lacking PTEN in some cells, by contrast, survive but spontaneously develop tumors in many organs (Bradley and Luo, 1998). Studies of heterozygous mice have revealed some unexpected genetic interactions, such as the dramatic decrease in tumors seen when they are crossed with PDK1 hypomorphic mice (Bayascas et al., 2005) and the dramatic increase when crossed with TSC2^+/– mice (Manning et al., 2005; Ma et al., 2005). Interestingly, a recent study demonstrated that the onset and incidence of tumor formation in mice lacking one copy of PTEN is highly dependent on the genetic background, even when the deletions are identical (Freeman et al., 2006). Differences in genetic backgrounds might well be why at least three phenotypically distinct syndromes [CS, BRRS and Lehrmitte Duclos syndrome (LDS)] arise from germline PTEN mutations, some of which are also identical (reviewed in Zbuk and Eng, 2007).

Generation of conditional PTEN-knockout mice showed that loss of PTEN results in increased size and number of cells in almost all cases (reviewed in Stiles et al., 2004). However, these models have also revealed the role of PTEN in additional diseases. Specific deletion of PTEN either from muscle (using MCK-Cre) or adipocytes (using aP2-Cre), for example, does not cause either rhabdomyosarcomas or lipomas, respectively (as might be expected). These mice are healthy and fertile, and are instead protected from high-fat-diet-(Wijesekara et al., 2005) or streptozotocin (Kurlawalla-Martinez et al., 2005)-induced diabetes. Adipocyte-specific deletion of PTEN further results in increased energy expenditure and body temperature (Komazawa et al., 2004). Loss of PTEN specifically in pancreatic β-cells also protects from streptozotocin-induced diabetes, although these mice are significantly smaller than control littermates (Stiles et al., 2006). Finally, PTEN^+/– mice are also protected from diabetes caused by knocking out insulin receptor substrate 2 (IRS2) (Kushner et al., 2005). Inhibition of PTEN in these tissues might thus represent an effective and non-toxic method for controlling type 2, diet-induced diabetes, a major world health problem.

Another non-tumor-related phenotype caused by PTEN deletion occurs following loss of PTEN in the neurons of the cerebral cortex and hippocampus. These mice exhibit a variety of abnormal social behaviors, increased response to sensory stimuli and decreased learning abilities (Kwon et al., 2006). Kwon et al. propose that this is reminiscent of autism spectrum disorders. Interestingly, a recent report found germline missense PTEN mutations in three out of 18 human subjects with autism spectrum disorders (Butler et al., 2005). Apart from macrocephaly, no other clinical features of CS, BRRS or LDS were seen in these patients, which further emphasizes the wide variety of phenotypes that result from PTEN alterations.

Attempts to associate PTEN alterations with additional neurological diseases, such as Alzheimers and Parkinsons disease, have been unsuccessful so far (Hamilton et al., 2006). Indirect evidence linking PTEN to Parkinsons disease (PD), however, does exist. PTEN-inducible kinase (PINK1, PARK6) was identified as one of 99 genes whose expression increases following expression of exogenous PTEN (Unoki and Nakamura, 2001). Germline mutations in PINK1 were subsequently shown to be associated with a rare form of early onset PD (Valente et al., 2004). Another gene altered in PD patients is DJ-1 (also known as PARK7). DJ-1 was independently isolated as a suppressor of PTEN function in a genetic screen in Drosophila (Kim et al., 2005). The exact mechanisms by which PINK1 and DJ-1 contribute to normal and pathogenic functions of PTEN remain to be determined.

Recently, an important role for PTEN in maintaining stem cells has emerged, which could affect how we interpret its tumor suppressor function and target it. Conditional deletion of PTEN in hematopoietic cells (Yilmaz et al., 2006; Zhang, J. et al., 2006) causes a myeloproliferative disorder and leukemia, which is consistent with its role as a tumor suppressor in many tissues. The number of hematopoietic stem cells initially expand, but ultimately decline, and cannot repopulate irradiated hosts when transplanted. Inhibition of PI3-kinase signaling by rapamycin simultaneously prevents leukemogenesis and restores the normal self-renewing capacity of the stem cell population. The inhibitory effect of PTEN loss on normal stem cell function might be tissue-specific, because deletion of PTEN in neural stem cells increases proliferation and the cells maintain their self-renewing capacity (Groszer et al., 2006). Deletion of PTEN in the basal stem cells of the prostate also causes increased proliferation, differentiation and prostatic intraepithelial neoplasia (Wang et al., 2006), which confirms the ability of PTEN loss in stem cell compartments to contribute to tumor formation. The effects of PI3-kinase-pathway inhibitors on tumor stem cells versus normal stem cells will clearly need to be studied before the safety profile of such drugs can be fully assessed.

The recent findings about PTEN regulation and function discussed above demonstrate that the pace of discovery is increasing, with the number of publications regarding this protein rising each year since its discovery. This has resulted in a greater understanding of the mechanisms underlying the regulation of PTEN and its roles in different human diseases, especially diabetes and cancer. Our comprehension of PTEN deregulation will hopefully translate into improved diagnosis, prognosis and treatment for patients. For example, PTEN levels can be used to stratify tumors that respond to new targeted therapies, including herceptin (Fujita et al., 2006), tarceva (Haas-Kogan et al., 2005) and tamoxifen (Shoman et al., 2005). Greater understanding of the different mechanisms underlying PTEN loss will undoubtedly refine the ability to more optimally treat patients. It will be exciting to see further studies examining the interaction of PTEN with other tumor suppressors or oncogenes. As new diabetes treatments are introduced, it will be interesting to see whether PTEN influences the response to these agents. Furthermore, as knowledge about PTEN post-translational regulation increases, more reagents should become available to examine under which conditions these occur. Another challenge for the future will be to understand the roles of nuclear versus cytoplasmic PTEN further, as well as phosphatase-dependent versus -independent functions and their biological relevance.

We would like to thank Gultekin Tamguney for critical reading of the manuscript.